From: Joel Sherrill <joel@rtems.org>
To: Jeff Johnston <jjohnstn@redhat.com>
Cc: Paul Zimmermann <Paul.Zimmermann@inria.fr>,
Newlib <newlib@sourceware.org>
Subject: Re: incorrectly rounded square root
Date: Fri, 4 Jun 2021 13:59:55 -0500 [thread overview]
Message-ID: <CAF9ehCW=3HNCnhdy80GvHfL88jvYW2LymfRMi0XquwRa2e8rcg@mail.gmail.com> (raw)
In-Reply-To: <CAOox84vMKMnqi1uGyO8tBO+5mjmnq_oAxqsqogCBz0c9Q--Rbw@mail.gmail.com>
On Fri, Jun 4, 2021, 1:44 PM Jeff Johnston <jjohnstn@redhat.com> wrote:
> Ok, I now know exactly what is happening.
>
> The compiler is optimizing out the rounding check in ef_sqrt.c, probably
> due to the operation using two constants.
>
> 86 ix += (m <<23);
> (gdb) list
> 81 else
> 82 q += (q&1);
>
> When I debug, it always does the else at line 81 without performing the
> one-tiny operation. The difference in the mxcsr
> register is the PE bit which I believe gets set when you do the one-tiny
> operation. Since we aren't doing it, it never gets
> set on and the difference of 0x20 in the mxcsr register is explained.
>
> By making the constants volatile, I am able to get the code working as it
> should. I have pushed a patch for this.
>
Awesome catch Paul and great eye to spot the problem Jeff!
--joel
>
> -- Jeff J.
>
> On Fri, Jun 4, 2021 at 3:14 AM Paul Zimmermann <Paul.Zimmermann@inria.fr>
> wrote:
>
>> Hi Jeff,
>>
>> > I figured the values were off when I had to hard-code them in my own
>> > test_sqrt.c but forgot to include that info in my note.
>> >
>> > Now, that said, using the code I attached earlier, I am seeing the exact
>> > values you are quoting above for glibc for the mxcsr register and the
>> round
>> > is working. Have your
>> > tried running that code?
>>
>> yes it works as expected, but it doesn't work with Newlib's fenv.h and
>> libm.a
>> (see below).
>>
>> > The mxcsr values you are seeing that are different are not due to the
>> > fesetround code. The code is shifting the round value 13 bits
>> > and for 3, that ends up being 0x6000. It is masking mxcsr with
>> 0xffff9fff
>> > first so when you start with 0x1fxx and end up with 0x7fxx, the code is
>> > doing what is supposed to do.
>> > The difference in values above is 0x20 (e.g. 0x7fa0 vs 0x7f80) which is
>> a
>> > bit in the last 2 hex digits which isn't touched by the code logic.
>>
>> here is how to reproduce the issue:
>>
>> tar xf newlib-4.1.0.tar.gz
>> cd newlib-4.1.0
>> mkdir build
>> cd build
>> ../configure --prefix=/tmp --disable-multilib --target=x86_64
>> make -j4
>> make install
>>
>> $ cat test_sqrt_2.c
>> #include <stdio.h>
>> #include <math.h>
>> #include <fenv.h>
>>
>> #ifdef NEWLIB
>> /* RedHat's libm claims:
>> undefined reference to `__errno' in j1f/y1f */
>> int errno;
>> int* __errno () { return &errno; }
>> #endif
>>
>> int main()
>> {
>> int rnd[4] = { FE_TONEAREST, FE_TOWARDZERO, FE_UPWARD, FE_DOWNWARD };
>> char Rnd[4] = "NZUD";
>> float x = 0x1.ff07fep+127f;
>> float y;
>> for (int i = 0; i < 4; i++)
>> {
>> unsigned short cw;
>> unsigned int mxcsr = 0;
>> fesetround (rnd[i]);
>> __asm__ volatile ("fnstcw %0" : "=m" (cw) : );
>> __asm__ volatile ("stmxcsr %0" : "=m" (mxcsr) : );
>> y = sqrtf (x);
>> printf ("RND%c: %a cw=%u mxcsr=%u\n", Rnd[i], y, cw, mxcsr);
>> }
>> }
>>
>> With GNU libc:
>> $ gcc -fno-builtin test_sqrt_2.c -lm
>> $ ./a.out
>> RNDN: 0x1.ff83fp+63 cw=895 mxcsr=8064
>> RNDZ: 0x1.ff83eep+63 cw=3967 mxcsr=32672
>> RNDU: 0x1.ff83fp+63 cw=2943 mxcsr=24480
>> RNDD: 0x1.ff83eep+63 cw=1919 mxcsr=16288
>>
>> With Newlib:
>> $ gcc -I/tmp/x86_64/include -DNEWLIB -fno-builtin test_sqrt_2.c
>> /tmp/libm.a
>> $ ./a.out
>> RNDN: 0x1.ff83fp+63 cw=895 mxcsr=8064
>> RNDZ: 0x1.ff83fp+63 cw=3967 mxcsr=32640
>> RNDU: 0x1.ff83fp+63 cw=2943 mxcsr=24448
>> RNDD: 0x1.ff83fp+63 cw=1919 mxcsr=16256
>>
>> Can you reproduce that on x86_64 Linux?
>>
>> Paul
>>
>>
next prev parent reply other threads:[~2021-06-04 19:00 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-05-04 8:08 Paul Zimmermann
2021-05-31 20:52 ` Jeff Johnston
2021-06-01 7:11 ` Paul Zimmermann
2021-06-01 16:28 ` Jeff Johnston
2021-06-02 7:51 ` Paul Zimmermann
2021-06-02 13:07 ` Joel Sherrill
2021-06-02 18:43 ` Jeff Johnston
2021-06-02 19:07 ` Marco Atzeri
2021-06-02 19:12 ` Joel Sherrill
2021-06-03 3:01 ` Jeff Johnston
2021-06-03 10:21 ` Paul Zimmermann
2021-06-03 15:50 ` Jeff Johnston
2021-06-04 7:14 ` Paul Zimmermann
2021-06-04 18:44 ` Jeff Johnston
2021-06-04 18:59 ` Joel Sherrill [this message]
2021-06-05 13:25 ` Brian Inglis
2021-06-07 9:51 ` Paul Zimmermann
2021-06-12 22:31 ` Maciej W. Rozycki
2021-06-03 11:25 ` Paul Zimmermann
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CAF9ehCW=3HNCnhdy80GvHfL88jvYW2LymfRMi0XquwRa2e8rcg@mail.gmail.com' \
--to=joel@rtems.org \
--cc=Paul.Zimmermann@inria.fr \
--cc=jjohnstn@redhat.com \
--cc=newlib@sourceware.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).